Nearly half of all recently published journal articles are freely available online, according to researchers who claim that institutions may be approaching a tipping point where they cancel their subscriptions to paywalled periodicals.
The suggests that about 19 million journal papers are now open access, and that these publications receive 18 per cent more citations than the world average.
Academics working in biomedicine and maths publish the most open access papers, while those working in chemistry, engineering and technology release the fewest, the report adds.
Heather Piwowar, co-founder of Impactstory, a non-profit organisation working to make scholarly research more available, led the analysis of 300,000 journal papers selected at random from three separate samples.
The first 100,000 papers in the sample came from the corpus of 66 million journal articles with digital object identifiers issued by the agency Crossref. Dr Piwowar and colleagues measured how many of these papers were open access using a tool that locates any freely available versions of journal articles.
“We estimate that at least 28 per cent of the scholarly literature is open access…Articles published in the last 20 years are increasingly open access, and this trend shows no sign of slowing,” they write in?PeerJ Preprints. For 2015, the most recent year analysed, 45 per cent of articles were open access.
Based on their research, Dr Piwowar and colleagues estimate that there are 18.6 million open access articles with Crossref DOIs currently available.
They found that the most common type of open access falls into a category that they dub “bronze open access”, and describes articles that are free to read on publishers’ pages, without an explicit open licence.
The researchers found that about 18 per cent of papers from 2015 within the sample were available as bronze open access. Green open access – articles published in a paid-for journal with an open access copy put in a repository, usually after an embargo period – made up a further 6 per cent of the papers, with 1 per cent featuring gold open access, which means that they are immediately freely available from a journal.
Dr Piwowar and colleagues garnered a second sample of 100,000 papers from a list of papers accessed by researchers over the course of a week in 2017 from a database of author-uploaded PDFs. A higher proportion of these, 47 per cent, were freely available, which they say is down to them being more recent.
Using a third 100,000-strong sample of papers sourced from the Web of Science, the authors were able to look at how open access varied by subject and how access affected citations.
“More than half of the publications in biomedical research and mathematics are freely available, while in chemistry and engineering and technology more than 80 per cent of the papers are hidden behind a paywall,” they write.
Researchers cited papers hidden behind a paywall 10 per cent less than the world average, but those that are freely available got 18 per cent more citations than the average, they added.
Dr Piwowar and colleagues say that the fact that about half of the recent literature is available without subscription could “tip the scales toward cancellation” of subscription deals for some universities.
“Increasingly, these libraries are under pressure to meet growing prices of ‘Big Deal’ subscription packages, and the once-unthinkable outcome of cancelling these Big Deals is becoming an increasingly realistic option,” they write.
